It’s common, if not almost required, for major transit projects in the U.S. to carry promises of traffic relief. In a practical sense they’d be foolish not to. Most Americans still drive from here to there, and winning the favor of the majority is the job of the politician. So we see all sorts of public transportation efforts tout the benefits they’ll have on congestion: from light rail to metro rail to bus-rapid transit to bike-share .

Construction on the second phase of L.A.’s Expo light rail line is expected to be finished this month, with an opening scheduled for 2016. The project comes with high hopes of reducing traffic congestion in America’s most notorious car town. That goal was outlined in the environmental analysis for the original leg of the line, from Downtown L.A. to Culver City, and was reiterated by Mayor Eric Garcetti this summer in discussing the extension to Santa Monica:

There’s just the little problem of the evidence. With few exceptions, studies tend to find limited signs that transit has much of an impact on nearby road congestion. Some places see slight congestion gains or mileage declines in the short term, and well-designed service should lay the foundation for reduced car-reliance in the long run, but the direct transit-traffic link is tenuous at best.

Blame induced demand

Part of the reason is that old devil of induced demand: the more space that opens up on a road, the more drivers emerge to grab it. (That’s also why building more roads doesn’t solve the traffic problem, either.) Another part is that public transit remains a marginal travel mode in most U.S. metros. That means a system can attract lots of new riders without moving the traffic needle much, especially if a big chunk of the new transit riders were also former transit riders—as opposed to people who gave up driving.

A new study of the Expo line spotlights those very explanations in reporting a modest link between light rail and congestion relief. Transport scholars Genevieve Giuliano, Sandip Chakrabarti, and Mohja Rhoads found the Expo rail to be responsible for an overall increase in transit ridership on the L.A. Metro system. But their data couldn’t detect any reductions in traffic speed or travel time on Interstate 10 in the same corridor, and they found “mixed and inconclusive evidence” for relief on local roads.

They conclude that the traffic benefits of the light rail (below, LRT) are “very limited,” and blame the surge of new drivers to new road space:

[T]he congestion reduction benefits of LRT are likely very limited. If new LRT lines are located in high demand areas, we can expect a lot of latent demand; so LRT—to the extent that transit capacity is increased—allows for new auto trips, some of which may fill up the space made available by those who shift from auto to transit (Downs 1992). Moreover, as in the Expo Line case, new LRT lines typically add only incrementally to the overall capacity of the corridor.

So it isn’t quite right for transit advocates or public officials to use “relieves traffic!” as a project’s main selling point. But it’s equally incorrect to use “doesn’t relieve traffic!” as a reason to oppose it.